Supporting Information
A combined 3D-QSAR and Molecular Docking Strategy to Understand the Binding Mechanism of V600EB-RAF Inhibitors
Zaheer Ul-Haq*, Uzma Mahmood and Sauleha Reza
Dr. Panjwani Center for Molecular Medicine and Drug Research, International Center for Chemical and Biological Sciences, University of Karachi, Karachi-75270, Pakistan
Supporting Information Number / Content / Page NoSa / Material and Methods / 1-3
S1 / The molecular structures and experimental inhibitory activities (pIC50) of V600EBRAF for all three groups A, B and C. / 4-22
S2 / Selection of the training and test sets by following the structural molecular diversity and whole range of inhibitory activities (pIC50) from the entire data sets. / 23
S3 / The actual and predicated pIC50 values with their residuals of the training and test set molecules generated by CoMFA and CoMSIA models for V600EBRAF. / 24-28
S4 / Docked confirmation of the most active compounds from the all three datasets: Compound 53 in magenta, yellow colour represent the Compound 04 and Compound 02 in blue colour from the Group A, B and C respectively. / 29
S5(a, b & c) / Docking interactions of three most active compounds from the three groups A, B & C within active site of V600EB-RAF receptor. These 2D interactions pose generated by the MOE software. / 30
Sa: Material and Methods
CoMFA studies
CoMFA derived by two descriptors fields known as steric and electrostatic. For the CoMFA descriptor fields 3D cubic lattice at different grid spacing 0.5, 1.5 & 2.0Å in all (x, y, z) directions were created to encompass the aligned molecules by Sybyl7.3. These descriptor fields were calculated using sp3 carbon probe atom with +1.0 charge and 1.52 Ǻ van der Waals radius to generate the steric (Lennard Jones potential) [1], and electrostatic ( Coulombic potential) fields with a distance dependent dielectric at each lattice point. The steric and electrostatic energy values were truncated at 30.0kcal/mol. The CoMFA-STD method in Sybyl was used to scale the CoMFA steric and electrostatic fields. However the best q2 model was obtained with a default grid space of 2.0 Å.
CoMSIA studies.
Klebe et al derived the Comparative Molecular Similarity Indices Analysis (CoMSIA) descriptors [2]. It is another form of 3D-QSAR studies which has reduced the some inherent deficiencies arising from functional forms of Lennard Jones and Coulombic potential in CoMFA technique. The CoMSIA method computes the hydrophobic, hydrogen bond donor and acceptor fields in addition to steric and electrostatic fields as used in CoMFA. CoMSIA descriptors were also derived with the same lattice box as used in CoMFA and all five fields were calculated with a grid spacing of 2.0 Ǻ employing a Carbon +1 probe atom with the radius of 1.0 Ǻ as implemented in Sybyl7.3. In CoMSIA, steric indices were related to the third power of atomic radii and electrostatic were derived from the atomic partial charges. Hydrophobic fields were derived from the atom based parameters; hydrogen bond donorsacceptors were obtained by rules based method. The attenuation factor was set at the default value of 0.3.
Partial Least Squares (PLS)
The correlation between CoMFA/CoMSIA molecular fields (independent variables) and the physicochemical property (dependent variable) of under consideration compounds were obtained by using PLS method [3-5]. For the current study pIC50 values were assigned as dependent variables in PLS regression analysis to drive models. Cross-validation (q2) analysis was performed by leave one out (LOO) method in which simultaneously one compound is omitted from the dataset and its activity is predicted using the model reproduced from the rest of the inhibitors which are the part of dataset. The PLS algorithm with the LOO method was employed to select the optimum number of components (ONC) and to assess the statistical significance of each model. The minimum sigma (α) was adjusted to 2.0kcal/mol as column filtering value and to improve the signal to noise ratio by omitting the lattice points whose energy variation was below the adjust threshold value. With the help of obtained ONC from the best selected q2valueperform the calculation of conventional correlation coefficient (r2). This value was used to assess the CoMFA and COMSIA models prediction ability, following equation used to calculate the r2pred. with respect to the test set.
(PRESS is the sum of the square deviations between the predicted and actual activity for each inhibitor in the test set, SD is the sum of the squared deviations between the experimental activity of test set and the mean activity of the training set.)
References
1.Kroemer RT, Hecht P (1995) Replacement of steric 6–12 potential-derived interaction energies by atom-based indicator variables in CoMFA leads to models of higher consistency. J. Comput. Aided Mol. Des. 9:205-212. DOI: 10.1007/BF00124452
2.Klebe G, Abraham U, Mietzner T (1994) Molecular similarity indices in a comparative analysis (CoMSIA) of drug molecules to correlate and predict their biological activity. J. Med. Chem 37:4130-4146. DOI: 10.1021/jm00050a010
3.Wold S, Ruhe A, Wold H, Dunn WJ (1984) The covariance problem in linear regression. The partial least squares (PLS) approach to generalized inverses. SIAM J. Sci. Stat. Comput. 5:735-743.
4.Clark M, Cramer Iii RD (1993) The probability of chance correlation using partial least squares (PLS). Quant. Struct. Act. Relat. 12:137-145. DOI: 10.1002/qsar.19930120205
5.Kettaneh N, Berglund A, Wold S (2005) PCA and PLS with very large data sets. Comput. Stat. Data Ana. 48:69-85. doi:10.1016/j.csda.2003.11.027
CAPTIONS OF SUPPLEMENTRY TABLES AND FIGURES
Table S1: The molecular structures and experimental inhibitory activities (pIC50) of V600EBRAF for all three groups A, B and C.
Figure S2: Selection of the training and test sets by following the structural molecular diversity and whole range of inhibitory activities (pIC50) from the entire data sets.
Table S3: The actual and predicated pIC50 values with their residuals of the training and test set molecules generated by CoMFA and CoMSIA models for V600EBRAF.
Figure S4: Docked confirmation of the most active compounds from the all three datasets: Compound 53 in magenta, yellow colour represent the Compound 04 and Compound 02 in blue colour from the Group A, B and C respectively.
Figure S5(a, b & c): Docking interactions of three most active compounds from the three groups A, B & C within active site of V600EB-RAF receptor. These 2D interactions pose generated by the MOE software.
1